QT PRODACT: sensitivity and specificity of the canine telemetry assay for detecting drug-induced QT interval prolongation.
The purpose of this investigation was to define the sensitivity and specificity of the canine telemetry assay for detecting drug-induced QT interval prolongation. Data from twelve studies generated in the QT PRODACT project were used in this investigation. The study design was a 4x4 Latin square cross-over design and included the following drugs: MK-499, E-4031, terfenadine, haloperidol, cisapride, bepridil, propranolol, diphenhydramine, captopril, verapamil, amoxicillin, and ciprofloxacin. The estimated root squared error of the model, which estimated the slope of the QT-RR relationships for each animal, for all dogs during the pre-dosing period was 5.45%. Using the QT-RR model, the sensitivity and specificity in each cutoff value that judges QT prolongation were estimated based on the experiment errors and measurement errors in the 12 studies. When the cutoff value was 5%, the sensitivity in 10% prolongation was 0.978 and the specificity in 0% was 0.996. In conclusion, it was judged that a 5% cutoff value for changes in heart rate corrected QT interval using the canine telemetry assay is practical, and the sensitivity and specificity of the telemetry assay are very high when using the analytical method presented here. Based upon this information, the canine telemetry assay using the individual subject heart rate correction model is recommended as a sensitive test system for the in vivo assessment of risk for QT interval prolongation.